Multi-objective surrogate-assisted optimization applied to patch antenna design

In this paper, we present a multi-objective optimization technique for designing a patch antenna on a finite ground plane by employing a surrogate model which is formulated by utilizing various regression and model selection techniques. The method presented enables simultaneous training and optimization of the surrogate model with few initial samples required. The techniques used enable faster convergence of the Pareto set when compared to a traditional multi-objective optimization scheme.